DisplayTitle

Predicting Protein Composition, Biochemical Properties, and Dough-Handling Properties of Hard Red Winter Wheat Flour by Near-Infrared Reflectance

¹USDA-ARS, Instrumentation and Sensing Laboratory, Building 303, BARC-East, Beltsville, MD 20705-2350. Names are necessary to report factually on available data; however, the USDA neither guarantees nor warrants the standard of the product, and the use of the name by the USDA implies no approval of the product to the exclusion of others that may also be suitable. ²Corresponding author. E-mail: sdelwiche@ asrr.arsusda.gov ³USDA-ARS, University of Nebraska, Department of Agronomy, Lincoln, NE.

Breadmaking quality in wheat is one of several considerations that plant breeders face when developing new cultivars. In routine breeding programs, quality is assessed by small-scale dough-handling and bake tests, and to some extent, by biochemical analysis of gluten proteins. An alternative, not yet fully examined, method for wheat flour quality assessment is near-infrared reflectance (NIR) spectrophotometry. The present study was performed on 30 genotypes of hard red winter wheat grown during two crop years at eight to nine locations in the Great Plains area of the United States. Biochemical testing consisted of measuring protein fractions from size-exclusion HPLC (M _r > 100k, M _r 25–100k, and M _r < 25k designated as glutenin, gliadins, and albumin and globulins, respectively), pentosan content, and SDS sedimentation volume. Dough-handling properties were measured on a mixograph and recorded as the time to peak dough development, the peak resistance, the width of the mixing curve, and the width of the curve at 2 min past peak. Partial least squares analyses on diffuse NIR spectra (1,100–2,498 nm) were developed for each constituent or property. When applied to a separate validation set, NIR models for glutenin content, gliadin content, SDS sedimentation volume, and mixograph peak resistance demonstrated reference vs. predicted correlations ranging from r = 0.87 to r = 0.94. Such models were considered sufficiently accurate for screening purposes in breeding programs. NIR spectra were responsive to each constituent or property at a level higher than expected from a correlation between the constituent or property and protein content (recognizing that protein content is modeled by NIR with high accuracy).